Sources and docs: Meta has dismantled its system for identifying and limiting viral hoaxes, and told content-ranking teams to stop penalizing misinformation
The company built effective systems to reduce the reach of fake news. Last week, it shut them down — I.
PlatformerCasey Newton
Context & Ripple Effects
This report supplies the operational detail behind Meta's early-January announcement that it would replace US fact-checking with Community Notes. It also extends an earlier retrenchment that included reduced support for WhatsApp fact-checking and CrowdTangle, suggesting the shift reaches both outside verification and internal distribution controls.
The change reverses the role of ranking as a backstop against false content: rather than only changing labels or reviewers, Meta is reportedly removing a mechanism that limited amplification. The later completion of the US fact-checking transition makes this report a key marker in Meta's broader moderation reset.
First-order effects
Meta's ranking teams are reportedly no longer to apply misinformation penalties, allowing content previously identified as misinformation or viral hoaxes to compete for distribution without that specific ranking restraint.
The reported dismantling immediately reduces the role of Meta's internal hoax-detection infrastructure in deciding what users see, while placing more weight on the replacement Community Notes approach.
Second-order effects
Fact-checking partners and other verification providers lose another route through which their assessments could affect reach, following Meta's earlier reduction in payments for WhatsApp fact-checking.
Creators and publishers whose posts were previously downranked for misinformation face a changed distribution environment; advertisers and users may have to judge whether the new system delivers comparable safeguards.
Third-order effects
If Meta sustains this approach, platform trust-and-safety policy shifts from proactive, centrally enforced ranking interventions toward post-publication, community-led context—a materially different model of the US fact-checking transition.
The outcome will test whether Community Notes can function as trust infrastructure at Meta's scale, or whether reduced ranking enforcement makes platform safety increasingly dependent on user participation and external scrutiny.
The trend: Meta's moderation reset is part of a broader move to replace centralized content enforcement with lighter-touch, community-mediated systems.
The fear that haunted me for the last couple of years was the notion that 2024 might be that last fair general election in US history. — I am even more afraid, now. [embedded post]
The new ‘anything goes’ Meta policies where misinformation demotion is dismantled does not fly well in Europe! #DigitalServicesAct #DSA — “For now, these changes only apply to the United States — though you get the sense that Zuckerberg would happily roll them out anywhere els…
This is why I never took a single penny of Facebook's filthy money. It was obvious even back in 2017 where they were going, and more obvious still in the intervening years [embedded post]
“after the company's own analyses found that these classifiers could reduce the reach of these hoaxes by more than 90 percent, Meta is shutting them off” [embedded post]
So: what's a Section 230 reform that *doesn't* come with a baby/bathwater repeal? I'm no fan of outright repeal, but ... Meta and the other platforms need to proverbially come to Jesus. [embedded post]
The OG “freedom of speech not freedom of reach” argument by Aza and me was articulated in this article I wrote in 2018. — Holds up incredibly well 6 years later... www.wired.com/story/free-s... [embedded post]
So wild to watch this drumbeat of Meta/Zuckerberg doing away with programs they once proudly touted as technical solutions to help safeguard democracy. [embedded post]
Get ready for the return of “Pope Endorses Donald Trump” and other similar viral bullshit. — This is part of the redefinition of “freedom of speech” to mean “freedom from moderation,” where any limit on algorithmic boosting or virality at all is ""censorship"". — www.platform…
Just in time for the inauguration and fascist shock and awe Zuckerberg shuts off all ability for meta to stop the coming tsunami of fascist mis and malign information. — People are not ready for what is coming with Trump 2.0s digital warfare campaign against the domestic popula…
“The result is that the sort of viral hoaxes that ran roughshod over the platform during the 2016 US presidential election — “Pope Francis endorses Trump,” Pizzagate, and all the rest — are now just as eligible for free amplification on Facebook, Instagram, and Threads as true st…
NEW: Meta has quietly dismantled the system that prevented misinformation from spreading in the United States. Machine-learning classifiers that once identified viral hoaxes and limited their reach have now been switched off, Platformer has learned www.platformer.news/meta-ends-…
I mean, based on his comments, it'll be replaced by Community Notes and Self-Reporting except in the case of clearly illegal stuff like CSAM. He's pretty much said he's going to lean more towards “free speech” even if more “bad” stuff gets in.
As far as I know (could be wrong) those classifiers were misidentiying some posts and incorrectly limiting their reach when they were not problematic. If you phrased this as “Meta is no longer using AI to determine which posts are limited in reach” people would probably say “Yay…